Research On Spectrum Sensing Technology In Cognitive Radio

As one of the key technologies of cognitive radio(CR), spectrum sensing becomes a current hotspot. Multi-user cooperative spectrum sensing(CSS) in CR network can better solve the problem of shadow effect and hidden terminal in actual wireless environment, and obtain higher detection performance of spectrum sensing. This paper further studies CSS and proposes improved algorithms of centralized and distributed CSS. Finally, the paper provides an energy saving strategy to select the sensing way of CSS. Then according to the strategy, CR users can make choices between centralized and distributed CSS.Double-threshold energy detection can be used as a local sensing scheme in centralized CSS. The common fusion method of Double-threshold cooperative spectrum sensing(DCSS) is that fusion center(FC) always needs to wait for merging until receiving all CR users’ sensing results and then makes the judgment. With the number of CR users increasing, the fusion method will need large channel resources to transmit sensing information. Aiming to solve the problem, this paper optimizes fusion method. Considering the fact that not all CR users’ sensing results are necessary when making the final judgment and the feature that each CR user sends their sensing results to FC by the means of time division multiplexing, in this paper, FC successively merges the sensing results and simultaneously makes the judgment depending on current-received sensing results. Then according to the judgment, FC makes choices between broadcasting sensing results of this cycle and merging the next CR user’s sensing result continuously. Simulation results show that the scheme of this paper can achieve better detection performance with smaller data overhead than DCSS.Since the present gradient based distributed CSS(GBDCSS) algorithms usually allow all CR users to participate in cooperative sensing activity, thus the participation of unreliable CR user will decrease the detection performance of CSS. Meanwhile, all CR users update their states equally in the present algorithms, failing to reflect the differences between each CR user, which affects the performance of CSS. To solve the above problems, this paper proposes an improved algorithm. This algorithm only allows CR users that hold better detection performance to participate in cooperative sensing, and uses credit to distinguish each CR user. This scheme can use credit to reduce the impact of CR users with poor detection performance and unsafe sensing results, then improving the detection performance of the whole CR system. Simulation results show the proposed algorithm can effectively improve the detection performance of spectrum sensing and combat attacks of malicious users.Since centralized and distributed CSS both have advantages and disadvantages, it’s difficult to decide which one is better. In order to solve the problem, this paper provides a solution from the point of energy saving. After balancing detection performance and energy consumption of CSS, CR users can make choices between DCSS and GBDCSS according to the proposed energy saving selection strategy.